Erratic behavior DIY CNC. Any suggestions on troubleshooting?

I'm experiencing erratic behavior. I can't figure out out. I've read everything I can find searching. I've wired and re-wired about 100 times. I get smooth motion in both directions randomly. Most of the time I experience smooth motion in one direction and noisy grindy sounding motion in the opposite direction. The direction doesn't seem to matter. If I swap the wiring to reverse direction of rotation, it's smooth the other way, then grinding and erratic back toward the direction that was previously smooth. Sometimes I'll be rough in all directions. When it's smooth it's great. I pretty easily get 600ipm, accurate, and then it just randomly goes back to erratic mode.

On the x axis it's completely smooth both directions with one driver and motor, then when I plug in the second it moves but one side will jerk and jump erratically - maybe once per 20 revolutions. Sometimes by turning on and off the computer I'll get different reactions from the motors.

I've wired with positive and negative pulse and direction. Makes no difference. I've moved around the motors and I have a spare motor. Makes no difference. I've also swapped drivers around. Again, no difference. I also have tried about every microstep option. Obviously speed changes, but otherwise about the same thing. I think it must be something related to the BOB or the computer - but how do I trouble shoot it? Any ideas on what I should look for? Any suggestions on why it works one way and not the other? I'm totally lost. I've tried everything I can think of.

Thanks!

Also the Mach3 Driver test shows excellent output.

Here's my CNC info:

My table is a belt drive x and y axis with ball screw z axis. 4x4 table. Not really important for this though.

I'm using a 4ish yr old Dell desktop with windows 7 and I have both mach3 and LinuxCNC. The same behavior shows up in both software. Unfortunately I don't own an oscilloscope, so unfortunately I can't really measure the parport activity.

I'm running two 36v 350w drivers. With a multimeter both have relatively stable output (looks stable on a multimeter).

The drivers are DQ42MA. I have 4 of them with 2 x axis motors.

Input voltage	18-50VDC
Input current	＜ 4A
Output current	1.0A～4.2A

Motors are Wantai 425in/oz Nema 23.

Part No.: WT57STH115-3004B/57BYGH115-003B

Frame Size: NEMA23

Step Angle: 1.8 degree

Voltage: 6.3VDC

Current: 3 A/phase

Resistance: 2.1Ohm/phase

Inductance: 9mH/phase

Holding torque: 30Kg-cm 425oz-in

Rotor inertia: 810 g-cm2

Detent torque: 0.89 kg-cm

Number of wire leads: 4

Weight: 1.55KG

Length: 115mm

And I'm using a db25-1205 BOB. It's powered with a 6v power adapter with 0.2A.

Quote:

---

Also the Mach3 Driver test shows excellent output.

---

Generally, the drivertest will show either excellent, or fail. So excellent is not always excellent?
Run it again, and see if it's a perfectly flat line. Also watch the pulse rate and see how much it varies. A really good PC won't vary by more than a few pulses. If it's more than 100 it's marginal.

What are your pulse widths set to in Mach3? Try setting them to 15 and see if it helps?

Are you running a direct drive to the belts?

If you run the machine at 50ipm, does it still do the same thing?

Those motors will not run very fast on 36V. They really need 100V or more to get the best performance from them.

Thanks for the suggestions. I'll go try pulse width variation. The test line is almost perfectly flat.

It acts the same regardless of speed, and it does the same thing on the bigger NEMA23s as well as the small NEMA 23 on the z axis with a ball screw. It also does the same thing on a spare motor that's not attached to anything.

I am running them direct drive. I chose high torque motors and run direct. Could be a bad choice. When it decides to work normally it's great. Plenty of power and very fast. If I could get it to operate all the time I'd be really happy.

Also is it possible that the BOB power source lacks sufficient amperage? That was one though I was kicking around was that it possibly wasn't enough power to the BOB and that was causing it to send weak pulses on the positive direction?

Since it does it with both Mach3 and LinuxCNC, my guess would be that it's something in the electronics.

Quote:

---

Originally Posted by ger21

Since it does it with both Mach3 and LinuxCNC, my guess would be that it's something in the electronics.

---

I agree. I have a degree in physics so I'm not totally new to electronics, but without any testing tools I'm kinda stuck with guess and check type diagnostics. And I'm totally new to stepper motors and CNC outside of commercial products.

My thoughts are the same as you on it being a hardware or maybe firmware problem. The problem is erratic, but displays the same behavior regardless of driver or motor. Unfortunately I only have one type of driver, but I do have 4 units and they all act the same. I do have 5 motors and 2 different models/brands. So I'm pretty confident it's not a motor issue. And I think it's unlikely that it's a driver.

Since it does the same thing in both software - but also runs normally sometimes in both software options - I doubt it is a software issue other than possibly the settings being wrong for the drivers. I think it's something related to the computer and/or breakout board.

My best guess at this point is either the computer parport is providing insufficient voltage or unreliable signals. It is a 2nd hand Dell optiplex. Or the bob is not transferring signals correctly.

The fact that it only happens in one direction seems to indicate to me that it is most likely a problem with voltage signal to the driver. My hunch is that the positive signals are not reliably providing enough voltage. What I can't explain is why it would be intermittent. Generally Ill get it dialed in, close up shop, then come back the next day and it's back to bad behavior.

Have you tried changing the acceleration values in motor tuning? You may be trying to drive them too hard. You might also try "Sherline half-pulse mode", if they really need a longer pulse . But like Ger says, those motors really want more juice - the high inductance takes a lot to overcome.

I've tried all sorts of speeds and acceleration values. Nothing changes.

Here's a couple videos. These are literally 10 minutes apart. No wiring changes. No changes in settings.
Good
<iframe width="854" height="480" src="https://www.youtube.com/embed/b8z7VXOcB6U" frameborder="0" allowfullscreen></iframe>

Bad
<iframe width="854" height="480" src="https://www.youtube.com/embed/YoizO7r_TP8" frameborder="0" allowfullscreen></iframe>

X axis:
<iframe width="854" height="480" src="https://www.youtube.com/embed/voBCwT1sD_k" frameborder="0" allowfullscreen></iframe>

Shortly after this I turned up the voltage to the BOB to see if it might help. Nope. It let out the magic smoke... So I'll be waiting on a new BOB in the mail. Not sure I can do much again until it arrives.

So, the obvious and total solution is for you to try my motor driver! LOL... sorry, but I have to try. Seriously, you might want to see if a different driver works, and the THB6064 kit I sell IS well thought of. Another option would be a geckodrive.com driver. For pete sake, do NOT buy a cheap chinese driver no matter what /they/ rate it at. Even the chinese THB6064's are junk... I have a video showing the difference.

Here are some other ideas:

1. Does it have the problem when only one axis is running, or is it more obvious when you are running 2 or 3 axis at the same time? If the latter, thin wires, or bad connections in the power supply wiring. Check the power supply at the driver, not at the power supply.

2. Does the problem happen only after it's been running for a while? Are your steppers getting hot? Everyone will tell you that "steppers are designed to run hot" and "it's no big deal" and I know beyond a shadow of a doubt that it IS a big deal, because they can lose 20% of their strength when they get hot. If it happens more when the motors are hot, put CPU coolers on them with thermal paste.

3. The Z axis is probably worse, no? And that is because it's the only axis battling gravity. It's not a "real" fix, but in marginal cases, a "quick fix" is to run a cable up over a pulley at the top and back down to a weight, offsetting the mass of the Z axis head. This means the z axis will now have twice the mass and will not move as fast, but it will move UP just as well as it moves down, which is the issue I'm seeing in your video.

4. Does the problem only happen on rapids? Higher voltage from the power supply should help that. If it happens all the time, more current might help... You could look into a battery based power supply with a charger.

5. It could easily be the optoisolators in your BOB. I hate optoisolators... try wiring a driver directly to the parallel port of the PC. If you fry it, the PC was old anyway right? I've seen bad optos cause problems just like this. Cheap crap from China again.

Hope some of that helps, and I'll make you an offer: Buy my THB6064 kit and try it, and post the results, and I'll refund the price of the kit.

I would try a different PC first.

Update:

I ordered a Sainsmart brand breakout board to replace the Wantai one that let out the smoke. Mostly because it was a 2day ship from Amazon. Put it in today. Everything works beautifully. All axis in all directions. Smooth, fast, and quiet.

I hope that this is the final solution. Seems to have solved the issues. I haven't set up a router on the table yet, but I'm very happy that this is solved.

Thank you for all the suggestions.

Hey, thanks for letting us know. I'll make a note to avoid Wantai BOBs and stick with the open source company designs (surprise, surprise).

:cough: it was probably the optoisolators :cough: ,o)

OK James. I just ordered one of your THB6064 boards hoping to cure my z-axis problem. If it cures the problem, I will order another three AFTER climbing to my roof top and singing loudly (and badly) your praises. I might even mention Lucas. Before you get too excited, I live in the country - no one will hear. Although, the cows may complain.

It sounds like CNCfrustration has solved his problem by smoking the culprit board, so I am risking a hijack of his thread. If I should do otherwise, let me know and I will comply as directed.

I built a Sidewinder CNC (V bearings on soft aluminium angle, springy plywood, but better than anything I had found at the time).

Lots of battles, things learned, things re-learned. I've been fooling around with this thing since last summer and still haven't gotten it to work reliably. When it comes close, I am thrilled.

I'm retired & cheap, so I tried to go the cheapest route and have paid twice for a lot of it.

My current problem is the Z-axis loses steps when moving up or down. This is random and whether it loses them on the up or down stroke depends on how many mini-bungie cords I use to offset the Z-axis weight. 7 is not enough, 8 is too much. With no bungies, it bores a hole through the basement floor. I have a spare motor, so swapped that in. No luck. I switched the X and Z drivers. No change. X and Y-A seem to work OK. I also switched the motors from the X and Z axes. Again a swing and a miss.

I can easily turn or stop any of the motors by grabbing the 3 inch handwheel on the back of each motor. It takes only light fingertip pressure. No pipe wrench required.

I'm getting tired of turning walnut into kindling.

CURRENT SETUP:
Universal Gcode Sender
Arduino UNO running GRBL .9i
CNC Shield (generic Chinese 4-axis)
Pololou 8825 drivers set to the maximum of 1.5 amps; set to single-step
Nema 23 Hybrid Stepper Motors: wired Bipolar Serial, 400 oz-in, rated current 2.83, resistance 1.6, VDC 4.52, inductance 7.2 mH
24 volt 20 amp DMiotech Chinese power supply adjusted up to 30 volts
4-start 8 tpi leadscrews; no backlash; very little, if any, turning resistance. This gives 1/2" travel per revolution. It goes like a scalded dog when cranked up, but I backed off to 50% on the X, 30% on the Y, and 20% on the Z.

Can your board be haywired into the CNC Shield in place of the Z-axis Pololou driver?

When I get the other 3 drivers, can I dispense with the CNC Shield and just plug directly into the Arduino pins? (I have a screw-connection shield)

When I get it all working, can I get another 24v power supply and put them in series to make 48v (dialed back to a safer 42v)?

Any help you can give an old fart with grand visions of a wondrous wood grinding machine will be gratefully, etc., etc., etc.

I have a couple of Arduino Mega's. What is a Ramps? I don't want to have to buy Mach 3 and a desktop to run it. V-Carve Desktop software is gonna be a big enough bite when I get a machine good enough to warrant it.

Other than mine, I have never seen a CNC Router in person, so all suggestions are welcome from any and all.

John

Quote:

---

Originally Posted by CNCfrustration

Update:

I ordered a Sainsmart brand breakout board to replace the Wantai one that let out the smoke. Mostly because it was a 2day ship from Amazon. Put it in today. Everything works beautifully. All axis in all directions. Smooth, fast, and quiet.

I hope that this is the final solution. Seems to have solved the issues. I haven't set up a router on the table yet, but I'm very happy that this is solved.

Thank you for all the suggestions.

---

Hmmm... you said you had a degree in physics, didn't you? I don't know how you connected the BoB which you converted to smoke generator, and I don't know which driver (there is no DQ42MA as far as I know) or the BoB you had or bought via Amazon, but I have two of these below, and both work just fine. I bought one from Wantai, together with five DQ542MA drivers and all work perfectly. On the other hand, I run the BoB with 5VDC/2A only, using the USB plug. It is not connected to my PC, instead I have a cheap USB charger as power source. What was the power supply type you used for the BoB? 6VDC 0.2A sound like a cheap crappy unregulated old NiCd charger. Is the BoB you have specified to be used with such voltage and VERY low current? If you fave used unregulated supply then the voltage is... unregulated, in worst case only through a rectifier and a tiny capacitor since 0.2A is pretty little and only needs a tiny capacitor to keep it somewhat stable. Those are NOT suitable for this kind of use.

Attachment 318640

I am still using 24V only for my motors and get 1500mm/min with the directly driven 1605 ball screws. Anything more than that and I get resonances/vibrations about half way down and I think that's normal. My DQ542MA can handle up to 50VDC and I will increase the supply to around 46VDC soon, but with your 36VDC you should not have any issues, at least to the same speed I get. Those drivers are very nice ones, never get hot, so you don't have to worry about anything other than your mechanical parts and the configuration. Don't listen to those who say you need 100VDC, that's nonsense, you will just burn those as well. Don't blame the Chinese, get some instruments and make sure you wire everything right and that the voltages are right and you don't have any earth issues and so on. Many things can cause smoke, so you must know what you are doing, not just guess and increase voltages. Don't blame the Chinese and believe that just because you buy it from Amazon you get better things. Even those can go up in flames if you don't know what you are doing.

My guess is that you have some kind of mechanical problems. Perhaps your couplings between the stepper and the ball screw. Perhaps you don't have a fixed ball screw end which causes severe vibration, perhaps you have far too high acceleration and a bad PC or PC card. Very difficult to say, but the least I'd worry about right now is the BoB and the drivers. Those are probably right, assuming you don't burn them and use them WITHIN the specifications, i.e. NOT connecting them to higher voltage than allowed AND have a power supply which provides enough current so that voltage drop due to load will not occur. You MUST use regulated supply for ALL electronics, unless there is a regulator in the BoB, but in that case 6VDC/0.2A is not enough, but there is an upper limit as well. For the motors you can use unregulated, but that's a different story. Maybe it is your configuration of the drivers. What is the microstepping you are using? How many "steps per", speed and acceleration and pulse width? Is it easy to manually turn the ball screw up and down without the stepper motor powered? How is the ball screw attached to the stepper? Is there a real fixture under that plate or is the screw attached directly to the stepper?

Post some pics of the gear. Makes it easier to guess.

Quote:

---

Originally Posted by OldSalt1945

OK James. I just ordered one of your THB6064 boards hoping to cure my z-axis problem. If it cures the problem, I will order another three AFTER climbing to my roof top and singing loudly (and badly) your praises. I might even mention Lucas. Before you get too excited, I live in the country - no one will hear. Although, the cows may complain.

---

I hear tell even the cows have internet these days, so as long as you post your experience (good or bad) I'll be happy. Feedback matters.

Quote:

---

Originally Posted by OldSalt1945

It sounds like CNCfrustration has solved his problem by smoking the culprit board, so I am risking a hijack of his thread. If I should do otherwise, let me know and I will comply as directed.

---

So, yeah, it's generally better to start your own thread, otherwise people get all confused about which symptom/setup was you and which was him. But since we are here....

Quote:

---

Originally Posted by OldSalt1945

I'm retired & cheap, so I tried to go the cheapest route and have paid twice for a lot of it.

---

And there it is... Now, mind there are two types of cheap... money cheap and time cheap. You can be one or the other, and NOT both. My kits are money cheap (but high quality components) but they need your time to assemble them. Hopefully the learning you did was worth the times you paid twice, in that it will save you time and money later.

Quote:

---

Originally Posted by OldSalt1945

My current problem is the Z-axis loses steps when moving up or down. This is random and whether it loses them on the up or down stroke depends on how many mini-bungie cords I use to offset the Z-axis weight. 7 is not enough, 8 is too much. With no bungies, it bores a hole through the basement floor. I have a spare motor, so swapped that in. No luck. I switched the X and Z drivers. No change. X and Y-A seem to work OK. I also switched the motors from the X and Z axes. Again a swing and a miss.

I can easily turn or stop any of the motors by grabbing the 3 inch handwheel on the back of each motor. It takes only light fingertip pressure. No pipe wrench required.

---

Classic case of weak motor / driver. In your case, I'd say weak driver because a NEMA 23 motor should be able to run any of that nicely. You may find that they start getting hot with the new drivers, but just add heatsinks to the motors (little CPU fans work great) via some heatsink compound and that will solve that.

Quote:

---

Originally Posted by OldSalt1945

Can your board be haywired into the CNC Shield in place of the Z-axis Pololou driver?

---

Yes, and because you are a nice person and will post a good review (:cough:) I'm going to throw in a free Pololu adapter for the driver kit you ordered. You could do it with just wires, but this makes it much nicer.

Quote:

---

Originally Posted by OldSalt1945

When I get the other 3 drivers, can I dispense with the CNC Shield and just plug directly into the Arduino pins? (I have a screw-connection shield)

---

Sure... anything that gets logic power, step, and direction to the driver is fine. If the cables are really long, you want shielding, but other than that, pretty much anything will work. We actually have an UNO to PMinMO (the connector standard we use is PMinMO) adapter but I've never bothered to document it and the pinouts might not match your setup. If it all works out and you order more drivers, and remind me that you want to try it, I'll toss one in free.

Quote:

---

Originally Posted by OldSalt1945

When I get it all working, can I get another 24v power supply and put them in series to make 48v (dialed back to a safer 42v)?

---

That's a whole 'nother conversation for another day. In general, sure... but you got to be careful putting standard power supplies in series to keep from letting out the magic smoke. Ground on the second one has to float. In general, not recommended. See this article for how to get away with it:
procooling com/index.php?func=articles&disp=52
and see this for general ideas on stepper motor power supplies:
techref.massmind.org/techref/io/stepper/power.htm

Quote:

---

Originally Posted by OldSalt1945

Any help you can give an old fart with grand visions of a wondrous wood grinding machine will be gratefully, etc., etc., etc.

I have a couple of Arduino Mega's. What is a Ramps? I don't want to have to buy Mach 3 and a desktop to run it. V-Carve Desktop software is gonna be a big enough bite when I get a machine good enough to warrant it.

Other than mine, I have never seen a CNC Router in person, so all suggestions are welcome from any and all.

John

---

RAMPS is a pretty standard "shield" for the Arduino Mega that is used in most 3D printers. It's pretty much a breakout board with Pololu connectors for the motor driver, terminals for the motors, relay drivers for heaters, terminals to connect temperature sensor, etc... It very much can be, and is on occasion, used to run a CNC machine. Again, just really a breakout board.

Lots of good advice on my site, linked in my signature.

Looking forward to hearing how you get on, John.

Quote:

---

Originally Posted by James Newton

Yes, and because you are a nice person and will post a good review (:cough:) I'm going to throw in a free Pololu adapter for the driver kit you ordered. You could do it with just wires, but this makes it much nicer.

---

Thanks for your quick response. Looking forward to receiving the THB6064 kit. I watched your 1:20:00 video, falling asleep only twice and saw nothing that put me off. I will wash the PCB with alcohol before starting assembly so I don't have the soldering difficulties you had. I will try to document the process and sing praises or engage in gnashing of teeth and vile curses as I deem appropriate. I should have it done within a day.

I will now get off of this thread and start a new one dedicated to this subject.

John